Forum Discussion
Monitor GEN2 Embedded capacity
I kinda lost here, so sorry for the amount of questions im asking here.
I recently started using a GEN2 A1 SKU for my embedded reports.
Right after, I set some metrics in my Azure Monitor and I installed the Premium Gen 2 Capacity Utilization Metrics app app.
First of all, what are the correct specs? This page and this page both list different RAM for instance. So what is it? 2.5 or 3 GB?
- What should I look at to monitor the load on the capacity?
- In the Azure Metrics app I can basically only monitor the CPU load and Overload. I’m guessing I need to monitor the max value? Or average? I’m not quite sure. For instance, the max CPU load shows CPU loads between 100% and 400% that does not sound like it makes sense? And as long as its below 100% its fine? Or what to look for?
- The overload that is supposed to show when the capacity reached its max? But so far every time someone logs in and checks a report the max overload is 1.
- Then the Premium Capacity Utilization And Metrics itself. Here it shows 0 overloaded minutes.
- What does the CPU graph in the Premium Capacity Utilization And Metrics say? It says it shows the CPU time in seconds, but how much should or shouldn’t that be?
- How to the graphs in the Azure Monitor work? (see attachment) The graph goes from 0 to 5m to 0.01?
3 Replies
Hi Tweeg
It is 2.5 because the two A1 equal an A2.
The following article will let you know which is more appropriate to detect the maximum value or the average value.
You can browse the entire article.
Best Regards
Community Support Team _ chenwu zhu
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
v-chenwuz-msft Thanks for the info! I have checked the page you linked but I still dont really get it.
Would someone be able to give a small (calculation) example or something?I agree with Tweeg, there is not RAM information on Power BI Gen 2 metrics in Azure Monitor nor Premium Gen 2 Capacity Utilization Metrics app. In our case we have serious dataset refresh errors for RAM consumption and can't fix it because we are "blinded", with no needed monitoring information.
What I didn't believe is to see this documented:
customers don't have visibility through utilization data to the memory footprint of their operations, and cannot know ahead of time whether any of their operations is subject to failures.
We need to solve these issues to publish reports in production but waiting MS support for many days, still counting...
